JC991 G46

INTERNATIONAL COUNCIL OF SCIENTIFIC UNIONS

I (

WORLD METEOROLOGICAL ORGANIZATION

THE GARP ATLANTIC TROPICAL EXPERIMENT <GATE)

REPORT ON

SALINITY INTERCot1PARISON

PELL ;.. ...... " .. J\.i~NCE liBRARY, Uf IV. of R. I.

. GENEVA, FEBRUARY 1976

P R E F A C E

This report is the f inal product of a cooperative inter­

comparison programme between countries making GATE oceanographic

measurements . Water sampl es were distributed to the parti cipating

countries before the GATE Experiment with the request that analyses

be made before and after the Experiment us i ng identical instruments

and procedures as in the routine sea water analysis. The results were

collected by the National Oceanographic Instrumentation Center (NOIC) in t he United States for analysis and intercomparison.

In order to aid the widespread usefulness of this report on

international intercomparison, the GARP Activities Office has agreed to

dis tribute these results among the national par ticipants in the GATE . I woul d like to thank Robert J . Farland of the NO IC for the preparati on of

this report .

David R. Rodenhuis GARP Activities Office World Meteorological Organization

QC997 G 40

SALINITY INTERCOMPARISON REPORT

* * *

The Oceanographic Sub-Programme

for the

GARP Atlantic Tropical Experiment (GATE)

Robert J . Farland

National Oceanographic Instrumentation Center

Washington, D. C.

November 1975

f t 11• ,* • ~.;~L ·: · '" -.,: r". • ·~o ~ nv 'JN!V oc ,, I ,, - . • • ,.,,.:

• , ., ~, •' . - , , .. L--1, )~"· •s ~~ ,.. · -· I r,,,,,t·v PAR:~GANSHi, R. I. 02882

APR-111975 .

GARP Act ivit ies Office WMO Secretar iat, C.P . No. 5 CH-121 I Geneva 20 Swi tzer land

Abstract . . . .

I. Introducti on .

C 0 N T E N T S

II. Intercomparison Reference Standards .

III. Qual i ty Assurance. . . . a . Solution P repa ration . . b. Bottling . . . . . . c . Quality Control Testing

IV . Intercompari son Resul t s. . . . v. Conclusion . . . . . . . .

Attachment 1 - Data code key

. . . . . . . .

.

PAGE

1

4

7

7

8

9

11

30

S A L I N I T Y I N T E R C 0 M P A R I S 0 N R E P 0 R T

The Oceanographic Sub-Prograr.tne for the GARP Atlantic Tropical Experiment (GATE)

Robert J. Farland National Oceanographic Instrumentation Center National Ocean Survey, NOAA, Washington, D.C.

ABSTRACT. Various valued salinity samples were provided by the National Oceanographic Instrumentation Center (NOIC) to 31 oceanographic vessels participating in the GARP Atlantic Tropical Experiment (GATE) during the summer of 1974. Intercomparison of the sample measure­ments by these vessels are to be used to evaluate the validity of the GATE salinity data prior to extensive scientific analysis. All the resultant measurements are referenced to NOIC's precision conductivity comparator standard. The intercomparison data revealed that vessels using the titration method had a standard deviation of 23.3 ppm, while those vessels with the inductive salina­meters (non-thermostated type) gave a standard deviation of 11.2 ppm. Some vessels used thermostated type salina­meters and the standard deviation for the data was 2.3 ppm. The data uncertainty level for the intercompar;son project was set at +10 ppm over a salinity range of 32 to 38 ppt. Repetitions of measurements are required to establish a more meaningful representation of the quality of the GATE data.

I. INTRODUCTION

Measurements of salinity during the 1974 GARP Atlantic Tropical Experiment (GATE) were obtained from many oceanographic vessels representing several countries using various types of instruments and methods. To evaluate the validity of these measurements, prior to extensive analysis of the GATE data, an intercomparison project was instituted with a goal for the data variance from a standard to be within +10 ppm. The project provided at least two sets, X andY, of various valued salt water samples (Table 1), prepared by the National Oceanographic Instrumentation Center (NOIC), to 31 GATE oceanographic vessels (Table 2), with salinometers or titration equipment (Table 3).

1

TAB LE 1. Sal t water sample sets and nominalize salinity values .

X Set Y Set SAMPLE NOMINAL SAMPLE NOfHNAL NUMBER SALINITY( ppt) NUMBER SAL! N ITY ( ppt)

XOl 38.0 Y07 38.0 - - --X02 36 . 5 Y08 36 .5 - - - -X03 35.0 Y09 35.0 -- --X04 33.5 YlO 33.5 -- --X05 32.0 Yll 32.0 -- --X06 36.5 Y12 33 .5 -- --

TABLE 2. Salt water sample sets distributed to GATE vessels.

DATA COUNTRY VESSEL SETS RECEIVED

Brazil SIRIUS 4 No SALDANHA 4 Yes

Canada QUADRA 4 Yes

France CAPRI CORNE 2 Yes CHARCOT 6 Yes

F.R.G. ANTON DOHRN 2 ( 1.) METEOR 2 ( 1.)

G.O.R . VON HUMBOLDT 2 Yes

Mexico MATAMOROS 4 No

Netherl ands ONVERSAAGD 4 Yes

United Kingdom DISCOVERY 4 Yes

2

TABLE 2. (Continued)

DATA COUNTRY VESSEL SETS RECEIVED

U.S.A. ATLANTIS II 4 Yes DALLAS 4 Yes H.J.W. FAY 2 Yes GILLISS 4 Yes OCEANOGRAPHER 6 Yes RESEARCHER 4 Yes TRIDENT 2 Yes

U.S.S.R. DEZHNEV 2 Yes KOROLOV 4 Yes KRENKEL 4 Yes KURCHATOV 2 Yes LOMONOSOV 2 Yes MUSSON 4 Yes OCEAN 4 Yes PAS SAT 2 Yes PORYV 6 Yes PRIBOY 4 Yes VIZE 4 No VOLNA 4 No ZUBOV 2 No

~TOTALS: 10 31 108

(l.)A data sheet was received from F.R.G. marked: METEOR PLANET DOHRN. Which vessel measured the sets \<Jas not identified.

(**)

TABLE 3. Salinity measuring equipment used in GATE intercomparison.

MANUFACTURER MODEL TYPE

NOIC PCC Inductive-thermostated Beckman RS-7B&C Inductive-nonthermostated Plessey 6230 Inductive-nonthermostated Auto Laboratory 601Mliii Inductive-nonthermostated Hydrometpribor Gt-165 Inductive-nonthermostated Guil dl ine 8400 Electrode-thermostated Woods Hole 3 Electrode-thermostated

(**) NOTE by GARP Activities Office: All samples from these FRG ships were analyzed by equipment in laboratories at Kiel.

3

Each vessel was instructed to determine the salinity of one set of samples before the start of a GATE phase and the other set at the conclusion of a phase . The .resultant data were recorded on data sheets (Example 1) which were transmitted to NOIC.

The nominal sa li nity values were repeated once in each set. For example, in the X set samples, the value 36.5 ppt was repeated and in Y set the value 33.5 ppt was repeated. Al so each set had a sample with a nominal value of 35. 0 ppt. It was assumed that each participating vessel would standardiLe their equipment by using a standard sea water of 35 .0 ppt. This scheme, therefore, provided repeatabi lity data at two points for each set duri ng the inter­comparison measurements : 36.5 and 35.0 for the X set and 33.5 and 35 .0 for the Y set.

All the water samples (more than 950) were initially measured by NOIC uti lizing its salinity reference standard system as the designated standard for this project. Approximately 25 percent of these samples were randomly selected and retained at NOIC to be measured for quality assurance monitoring purposes. The measure­ment schedule was designed to be carried out before, during and after the at sea portion of GATE . Also, several of these samples were mea sured by other national and international laboratories, recogn ized for their quality salinity measuring capabilities. The resultant data provided a quality control check on the performance of the NOIC salinity standard. The overall flow of the various intercomparison salinity measurements for GATE is shown in Chart 1 .

I I. INTERCOMPARISON REFERENCE STANDARDS

Several standards were referenced throughout the salinity intercomparison project. In al l of NOIC's measurements~ the traceable chain was from its instrumental standard, the Precision Conductivity Comparator (PCC), to I.A.P.S.O. standard sea water and eventually the Unesco oceanographic tables for the final salinity readings.

The PCC is a self-contained thermostated system, that compares the conductivity ratio of an unknown sample with that of a known reference standard. The conductivity measuring circuit consists of an oscillator, a precision ratio transformer, a null detector and six conductivity cells, each 40 cc, one cell being used for the standard water, and the other five for the water under test. The cells are fabricated in quartz and immersed in a 26.5 litre temperature-controlled oil bath. A specially adapted inductive bri dge is employed with a platinum thermometer in the bath, to moni tor the temperature stability of oil surrounding the cells to

4

E)

g-1.

2.

cour

3.

4.

f-10(

L

t'

J

I

1.

2.

EXAMPLE 1

PROG~Ai1 NAME

I ~TERCALI BRAT I ON (INCAL)

SALINITY DATA SHEE T

PARTICIPANT'S ORGANIZATIONAL NAME .

COUNTRY VESSEL MEASURER'S NAME

3. TYPE OF SALINITY 0 INDUCTIVE SALINOMETER I 0 OTHER r~EASUREMENT (CHECK ONE) CJ TITRATION {METHOD & EQUIPMENT)

4. INSTRUf~ENT MANUFACTURER

HODEL NUMBER I SERIAL OR IDENTITY NUMBER

5. STANDARD FOR SALINITY MEASUREMENT (CHECK ONE)

CJ. IAPSO STANDARD SEA WATER BATCH NUMBER I DATE

0 OTHER (EXPLAIN)

6. TABLE OR EQUATION USED TO COMPUTE SALINITY (CHECK ONE)

0 INTERNATIONAL OCEANOGRAPHIC TABLES (UNESCO)

0 OTHER (EXPLAIN)

7. SALINITY DATA GROUP X MEASUREMENT DATE

TEST SAMPLE TIME SALINITY OF

:

IO NUMBER (LOCAL) TEST SAMPLE (ppt) 1 2 3 4 5 6

GROUP Y MEASUREMENT DATE 7 8 9 10 11 12

8. ADDITIONAL DATA (ADDITIONAL SPACE ON REVERSE SIDE)

t ..,.. .. w "*' caull'ht»4 -• m - aMMo nm~t. - ,... ,

I

'

-

. i ~

CHART 1. Flow of the GATE Salinity Intercomparison Measurements

TITRATION DATA

FIELD GATE PARTICIPANT NON -THEW40STATED SAMPLES MEASUREMENTS SALI NOMETER DATA

THERMOSTATED SALINOMETER DATA

CONTROL/STANDARD THERMOSTATED SAfi.PLE LOT

MEASUREMENT SALINOMETER DATA PREPARATION

AT NOIC ENVIRONMENTAL (Q.A.) THERMOSTATED TESTS & MEASUREMENTS SALINOMETER DATA

I

TITRATION DATA

Lt.

CmlTROL COLLABORATIVE NON-THERMOSTATED SAMPLES MEASUREMENTS SALINOMETER DATA

- THE RMOSTATED SALINOMETER DATA

6

~0.2 mK. The dimensional stability of quartz incorporated in the design of the test cells effectively minimized changes in their cell constant with time and temperature. The PCC facilitates salinity measurements from 4) to 60 ppt over a temperature range from -5°C to 40°C with an uncertainty of 1 ppm.

Data obtained from other laboratories were referenced to such instrumental standards as: Guildline Salinometer, Model 8400 (Bedford Institute of Oceanography, Canada); Autolab Model 601, Mk. III (Institute of Oceanographic Sciences, England); and WHOI Salinometer, Model 3 (Woods Hole Oceanographic Institution, USA). Generally, salinity measurements from inductive type salinometer~ \•lere traced to I .A.P .S.O. standard seawater and the Unesco oceanographic tables.

During t he intercomparison project , information received from the participants indicated that other standards were used in their traceability of salinity measurements. The Canadians utilized the conductivity-salinity tables of A.S. Bennett, Bedford Institute of Oceanography. A few of the Soviet vessels traced their measure­ments to standard seawater prepared by the Institute of Oceanology (USSR Academy of Sciences) and theN. N. Zubov (Leningrad 1957) tables.

In anticipation that some vessels would measure the samples by the titration method, a minority of NOIC's quality control samples were titrated by the U.S. Naval Oceanographic Office (Washington, D.C.) who used the Knudsen method \'lith a phenosafranim indicator as described in their publication entitled: "Instruction Manual for Oceanographic Observations 11 No. 607, 1968.

III. QUALITY ASSURANCE

a. Solution Preparation

The samples, represented by sets XOl through X06, YlO and Yl2, were prepared from natural seawater obtained from the Sargasso Sea area with a salinity range of 36 to 38 parts per thousand. This seawater was concentrated to approximately 40 ppt by evaporatiQn at 40°C. For sample sets Y07 through Y09 and Yll, the solution was constructed synthetically by the use of a commercial salt known as Sea Salt. The salts were dissolved by weight in deionized water, producing a solution of about 45 ppt. The absolute solution value was not critical at this time. The solution was al lowed to settle for a week and then the clear salt water was drawn off and filtered.

7

All the concentrates were filtered as they entered a 50 litre mixing and dispensing bottle. The deionized water was added to dilute to the desired nominal salinity. A partial vacuum was drawn for about a half hour on the solution while stirring with a magnetic stirrer to degas well below saturation at room temperature . Atmospheric pressure was restored and the solution constantly stirred until half-way through the bottling.

b. Bottling

The sample bottles used in the intercomparison were all borosili cate type at 650 ml size. Soda-lime glass bottles \'t'ere first considered because of their low cost and easy accessibility. These bottles would then be acid rinsed (10% HCL) and subsequently washed wi th distilled water. Soda-lime glass (or flint glass) has approximately four times more soda (Na 2C0 3 ) than borosilicate glass . The soda will leach out from the glass surface at a rate decreasing with ti me. Table 4 is an example of test results using distill ed water in a 10 oz . soda-lime container which was heated to 120°C for 1 hour . This is equivalent to 1 to 2 years on the shelf at ambient temperature.

TABLE 4 . Reactions of soda-lime containers, distilled . water and heat.

OXIDES QUANTITY (ppm)

Si02 26.0

Al203 0.6

GaO 4.1

MgO 0.4

Na20 6.5

K20 0. 01

Bz03 0.01 -

These oxides will increase by 2 to 5 times if the pH of the water were 8.4. Alkalinity therefore accelerates the leaching process .

8

To assure quasi absolute non-contamination and stability of the water sample composition, the borosilicate bottle was chosen over the cheaper soda-lime glass container. Measurement integrity was of prime importance.

The sample bottles were prepared prior to filling with the desired solution by rinsing in deionized water, drying in an oven for 30 minutes at a temperature of l50°C, then allowed to cool in an inverted position in a drying rack.

The bottles were readied to receive the various valued salinities. The solution was drawn off the bottom of the 50 litre mixing and dispensing bottle via a glass dip tube (the only entrance to the bottle except for a small air vent at the top) and gravity fed via tygon tubing to the sample bottle. The tygon tubing was flexible enough to stop the solution flow by pinching the tubing by hand. This eliminated spillage of water on the sealing surface of the bottles. The sample bottles were then sealed with b1o layers of a highly flexible thermoplastic, known as parafilm, and a layer of aluminum foil. Parafilm is waterproof and prevents sample evaporation and contamination. The aluminum foil allowed the cap to be screwed on with no pulling of the parafilm. The cap was tightened down snugly with a pair of pliers, and the bottle labeled. The first and last samples bottled for each salinity value \·Jere measured to check for any differences. The standard deviation for the twelve salinities was 0.6 ppm.

c. Quality Control Testing

A control sample lot of the bottled seawater was retained by NOIC and measured over a period from April to November 1974 on the PCC. The dates were: April 2 and 5; June 7; July 15 and 16; August 7, 20 and 28; September 17; October 1 and 4; and November 27. Twenty sets (120 measurements) were tested during this period with results in Table 5.

The results of the control samples measured by other laboratories are shown in Table 6.

9

TABLE 5. Measurement results of NOIC control salinity samples.

SAMPLE PCC MEAN PCC STD. DEV.a NUMBER ( ppt) (ppm)

XOl 37. 9403 1.1 - -X02 36,6508 1.1 --X03 35.0549 0.8 --X04 33.5099 0.6 --xos 31 .9754 1.0 --X06 36.5244 1.1 --Y07 38.0475 1.4 --Y08 36.6826 1.2 --Y09 35.0227 1.5 --

YlO 33.5631 1.0 --

Yll 32.0616 0.9 --Yl 2 33.5235 0.9 --

TABLE 6. Salinity differences between NOIC 1s PCC and other laboratories.

SALINITY DIFFERENCES (ppm) Nomi na1 Sa 1 inity

(ppt} 32.0 33.5 35.0 36.5

Sample Group xos Y11 X04 YlO Y12 X03 Y09 X02 X06 Y08 Number Data Code Identity

611 0 _,

1 BB 1 2 1 0 0 1 615 0 2 1 1 0 1 3 0 1 3 616 12 11 10 20 12 8 11 6 9 11

Data Key: BB-Broken Sample Bottle; no data .

10

38.0

X01 Y07

1 1 l 3 6 9

Simula ted environmental tests were also performed on the samples to determine if any measurement degrading factors coul d be at tributabl e to shi pment by air or storage at sea . The samples were tested under as sumed conditions most likely to occur during trans portation and subsequently on-board a GATE vessel . Environ­mental temperature, pressure and vibration were the prime parameters .

The test samples were placed in an environmental test chamber where the ambient pressure was lowered and the temperature was made to range from 32°C to l0°C . A visual investigation of the samples in a bright light indicated no change or crystal lization of the solution . The samples were then heated to, and maintained at a temperature of 50°C for 6 hours. This test was to simulate the samples being on a ship's deck in direct sunlight. The samples were cooled to l°C for 3 hours. At this temperature the samples were allowed after the waiting period to drift overnight up to l6°C and then recooled to l0°C under an ambient pressure of 585 mmHg. The samples were also exposed to vibration over a frequency range of 20 to 58 hertz at a temperature of 50°C . This frequency range was selected t o duplicate transportation vibration . Agai n, visual inspection of these sealed samples (simi l ar t o t hose del i vered t o the GATE vessels) indicated no discernible change .

All the test samples , following the environmental t ests, were remeasured during November 1974 with the PCC to determine the degree of salinity change . The maximum difference in salini ty was 4 ppm, but the average difference was less than 2 ppm.

IV . INTERCOMPARISON RESULTS

The t otal measuremen ts from GATE vessels were 560 data points with approximately 57 percent of this data fall i ng with in the +1 0 ppm salin i ty goal of the intercomparison project . The average standard deviation for this group was 10.4 ppm. Table 7 is an analysis of the three types of measurements used from t he GATE vessels. Data resulting from the titration of samples was done generally as referred to earlier in H. 0. Publication 607, "Instruction Manual for Obtaining Oceanographi c Data". The exact method used by each GATE partici pant was not requested and sub­sequently not detailed herein. The measurement differences experienced by titration was typically higher than those by other measurement techniques, as only 31 percent of the measurements were within the +10 ppm region . Other measurements (non-titration) were grouped into two categories: Measurement by thermostated salinometers and measurement by non-thermostated salinometers. It can be noted that 65 percent of the measurement differences from t he non-titration group fell within the +10 ppm region.

ll

Thermostated sa1inometers are those such as the Gui 1dline 8400 and WHO! Model 3, that uti li ze. a temperature controlled bath whi ch maintains both the reference sampl e (standard sea water for exampl e) and the unknown sample at t he same temperature to measure t he con­duct ivity ratio . The data observed from these type devices, in general, showed a significantly smal l er measurement difference and standard deviations as i l lustrated in the data of Table 7. In comparison, the non-thermostated salinometers showed measurement differences and standard deviations quite lower than that of titration, but somewhat higher than that of thermostated sali nometers . The non-thermostated salinometer was observed to have 61 percent of its data with the +10 ppm range as compared to 100 percent for thermostated salinometers. The non-thermostated salinometers (Plessey , Hydrometpribor GM-65, Auto Lab 60k, MK III and Beckman RS-7) measure the conductivity ratio between the unknown and reference sample which are not necessarily at the same temperature. To compensate for this temperature difference, an electronic t emperature compensating circuit is utilized. All available data indicated that the Hydrometpribor GM-65 salinometer was of the non- t hermostated type and, therefore its data was treated as such .

TABLE 7. Analysis of measurements from GATE vessels.

QUANTITY DATA OBSERVED STANDARD OF WITHIN +l 0 ppm DEVIATION OF

MEASUREMENT TYPE MEASUREMENTS (PERCENT) TYPE (ppm)

Titration 143 31 23.3

Non-Thermostated 370 61 11.2

Thermostated 47 100 2.3

The data received from the GATE vessels has been compared to the project's standard, NOIC's PCC . The results are tabulated in Table 8. The measurement differences were formulated using the raw dat a contained in the data sheets from the GATE vessels. Nominal sa l inities are noted at the top of the table, in ascending order from left to right . Each horizontal data code shows the measurement differences in ppm between the (NOIC) PCC's established values and the GATE measurer's readings.

12

TABLE 8. Salinity differences between NOIC PCC and GATE vessels.

SALIN ITY DIFFERENCES (ppm)

NOMINAL SALINITY 32.0 33.5 35 .0 36. 5 38.0

(ppt)

SAMPLE GROUP X05 Y11 X04 Y10 Y12 X03 Y09 X02 X06 Y08 XOl Y07 NUMBER

DATA CODE IDENTITY

617 7 -4 9 -9 -9 15 -14 14 5 -22 22 -17 618 -5 -4 -1 -4 NO -9 -3 -9 -3 -3 -8 -4 619 -2 -1 -3 -1 -1 -3 0 -5 -1 0 -4 -1 620 -12 2 -13 0 -6 -35 -8 -38 -39 - 18 -38 -20 621 28 25 27 16 17 12 -2 10 4 -13 -1 6 -20 622 6 9 6 8 8 3 2 3 -2 1 2 -2 623 6 9 1 7 6 -9 -1 -14 -12 -4 -19 -6 624 10 18 6 NO 7 -2 7 -3 -4 -2 -8 3 625 2 4 1 1 2 -1 1 -2 -5 0 -2 -1 626 3 2 2 2 0 0 -2 -2 -3 -2 - 5 0 627 -1 -10 -5 -20 -23 -10 -14 -13 -12 -24 -23 -34 628 10 21 17 15 14 2 10 8 10 10 2 14 629 22 33 31 32 21 -7 18 19 10 21 11 14 630 -13 17 -26 5 5 -33 0 -7 -10 28 -18 21 631 -4 -10 -14 -13 -22 -19 - 18 n 0 7 -1 -3 632 31 29 28 16 18 15 12 27 14 7 19 1 633 24 5 13 3 0 8 1 3 3 -9 2 -18 634 -24 -10 5 -21 - 12 -24 1 -29 -21 -21 -27 -1 4 635 -85 -32 - 30 -13 -13 -25 -13 -31 -74 -53 -20 -97 636 -15 -12 0 -33 -23 -5 -3 -31 -14 -43 -20 -27 637 -15 -12 0 7 -13 -25 7 19 -14 7 -20 - 7 638 5 8 0 -1 3 -33 -5 7 19 6 7 -20 -27 639 -15 -12 0 -13 -13 -45 -13 -11 -14 -33 0 33 640 -10 2 -6 -27 * -38 -3 -5 -18 -17 -3 23 641 -35 8 0 -53 27 -45 27 - 51 -1 4 -33 -20 -7 642 -30 10 -7 -59 17 -36 31 -40 0 -17 -32 -3 643 -15 -62 0 -33 -13 -25 -13 - 11 26 -13 50 33 644 -19 -51 -2 -44 -21 -19 -2 -14 26 3 43 27 645 9 23 18 32 34 -4 3 11 15 14 23 31 646 -13 23 20 31 27 -6 -11 8 15 3 17 16

13

TABLE 8. (Continued)

SALINITY DI FFERENCES (ppm)

NOMI NA L SALINITY 32 .0 33.5 35 .0 36 .5 38.0

(ppt)

SAMPLE GROUP X05 Yll X04 YlO Y12 X03 Y09 X02 X06 Y08 XOl Y07 NUMBER

DATA CO DE IDE NTITY

647 -15 28 20 7 27 -5 27 39 6 47 0 53 648 5 8 20 7 7 -25 -13 19 6 7 0 13 649 23 23 15 16 1l 8 6 -3 -7 -2 -15 -17 650 21 19 14 13 10 9 3 -4 -13 -2 -14 -14 651 27 19 20 13 12 8 10 3 -9 -2 -12 -9 652 30 3 32 -3 -1 8 16 24 7 16 -6 -13 653 9 8 0 2 3 -5 2 -8 -7 -6 -10 -8 654 14 15 8 8 9 2 5 -5 -2 -1 -6 -7 655 16 6 12 13 2 6 0 * 4 2 -2 -2 656 6 -59 10 3 11 14 0 18 -8 9 10 -8 657 -5 0 1 -6 0 24 -15 3 8 -7 -3 -10 658 48 28 3 51 49 0 5 -2 45 -6 -5 -6 659 1 -2 2 -5 -4 1 -7 -5 -6 -8 -1 -9 660 3 -2 2 -2 -2 0 -1 1 0 -5 1 -7 661 2 3 -2 2 1 0 3 6 1 2 -1 1 662 4 2 1 2 l 1 2 -1 0 1 3 0 663 3 1 1 0 0 -3 0 -3 -5 -2 -5 -6

Data Key: NO - No data received for the sample * - Unrealistic data; difference greater than 99 ppm

Each GATE vessel, that measured the X andY samples, received an approximate repeated value in those groups. The nominal salinities of 33.5 ppt and 36.5 ppt were the repeated values. Also a nominal sali nity value of 35.0 was provided in each X andY sample group. It was assumed that the GATE vessels would exercise generc1lly accepted operational practice with their equipment by standardizing at the standard of 35 ppt. This action, therefore, would have the effect of repeating the 35 ppt value when measuring the X and Y samples.

14

Table 9 is an analysis of these repeated measurements. The amount of measurement differences from the PCC reference standard to be within ~1 0 ppm is shown in percent.

TABLE 9. Ana lys i s of the repeated sampl e values measured by the GATE vessels.

OBSERVATIONS WITHIN +10 ppm (Percent) -

NOMINAL 1• SALIUITIES 33.5 35.0 REPEATED (ppt)

EQUIPMENT

Titration 17 25/25

Guildline 100 100/100

WHO! 100 100/100

Plessey 86 86/100

Beckman 85 57/71

Auto Lab 50 100/75

GM 65 27 69/62

1 ,

36. 5

21

100

100

85

64

63

62

Percentages represent X/Y sample groups; no repeat samples for this value were provided ; it was assumed that the GATE . vess~ls stan­dardize thei r equ ipment at 35 ppt with standard wa t er.

Al l comparisons, whi ch include the GATE vessels, other laboratories and the quality control checks on the PCC, are graphically shown in Figures 1 through 13. The curves for each plot represent data received from measurements on the X and Y sets of samples . The hatched area is the ~10 ppm sa l inity t olerance level desired for all the GATE salinity data . The graphs have been grouped in accordance with the method used and t he instrument: Titration ; electrode salinometer (Guildline and WHO!); and inductive salinometer (PCC, Plessey, Beckman, Auto Lab and GM65).

15

I ., . : : ! t

"" ··- · -L .. ___ _i _ . .. ... . l_. ____ _! _" . ...

i : I I $0

I l I

L __ _ L_.- ~

i~ 3'/ 3' ~8 WI

; . .S:LIN/TYc;1J. : I

.. I ... .. _! __ - -

I

i .3'1- 3~ .$8 <,0

5 ALINITYV/i)

FIGURE 1 •r

i .. .. • i

1t- ··--

t , I I : .

i I

I

..1tJ I

I

l

' I .JI/ X .38 ~

S AL/AIITY. (//t') ·- !

----.

I

I -- I

~ I

I

I__- --· ··-

' '

' I

-63~-

3~ 3~ ~d

sAL.IN!TYc;,?t) ! • !

! I

I I

- -. ~ ---·----- i I

-· j -.

I

- ! ... -'

·--! ,._ -- -- ·

FIGURE 2 , ..,

-638f­l !

···-·!

I I

I I I I I

*' I

S ALINITY l'ivr£ RcoHPAI?I.SON

(TITRAT/~N) I I

I I - i !~( SA(1PLEf Ty SAHPL1S I

~ .. ,__ _;'17- !-- ~--· - 1- I +,-- f -·-

~ 'fO ll • ; ! - 6 Lf8,-'J I I

~ .. -1----+ - 2# --- - !-- L___l ~ loJ--~.,..----,~~..,.---1-..,..-~...:----=- : ;o r--.::-~' *"---'~:""--.::~-~ """"=~~~..---1 ~ (> t--.......... :--~::----"'\--1-~-,.,...__..::~ l Crt

I

-1 I

GATE I j

I I l M~~~~-~~~-~--~~ -~- •. .2o _______ zo __ a;L. __ J'l . Jf _ .1$_·--·~ --

1- . I

I I j ___ _

I

I· I

I

. SA:LIIVIT'(~/'i)t : i ! ! I

I I

.3'1 3C. 38

s1LINITYifl't,> i

_-_ 661-

I

I 3()

I I !

i ! . I

I I

I i I I

· I!:J

___ .32 -·· . J'l ~ ~ . -- --~- -- · · -·1 s 1LIN/TYv/'kJ ; ,

l I I : I • • I

· r

3'f .Jr.- $~

SALIN/ TV(//' t } i - ·- I · I • j : ; • i

I I

I ;

_-_66e~

. 311 -·· 9~ 'e SJ!LI N / '[ Y ifJ~t)

I I

I \

-·--·-· :

- --··· -· I

i . I i _ .. 4_.J._; ____ ~ ·----~~-l __ _

' I I

·-· ---· ·-· -

FIGURE 4 19

I ; ! ~ - -·· .. ____ j ___ ·- .. .. --- -~ .J

I

GLlTE SALINITY J7NT;RCON'p4/?/SON-' 1 I I I

~ ---_ -~-----. ~~ P"NDt/Clt'E -1AL/Air:"fEr£R-~c) _ :-. - ~---:- __

! I "' )( SANPLks ~y SA1'1PteS 1 l

l , I I -~ ~ I I -60/- ;20 - 60'-1-1 ~ 2o I I ~ lo I . I /0 ........,. __ --~..~--:.:--~---.:,-----.,--1

~ 0 ---1- -·-· . 0 k._-=::::·~~=ll!P:;;;=l!~;;;$..::::...,~

~ I , ; ~ ...., __ : __ Jk. 3¥ ' )~ - J6

~ i SA~INITY~/"t) !

..j .

i /()

.,0 1 ;i!D . I I I : i I I i I

' I

I I 31/ ... ~{. .. . ~

SAL./Nl~Y ~j>t) I - _1!_ - ·- - I ~ -- -·· -- l . ~- i .

I . ___ i. ·-

!

I .. ··- ~ --~+ .. -·-·· -l - ··--·

I

-60C.-

' --· - -·- · - - .. - --·r. -- ·- ···i ·~ -··

I __ _ I

I I

___ _I

I

l l i

I I l

--- I ..... I

i i_

FIGURE 5

. I -· . I

-605+ I ' - · ·---~ ·-· - -.. -·-: -·

i

~~--~~--~--~~-~~ --~

l ' . - - .1¥ ..... -~- :18

SA/. I All T((/>;>t')' I : . I

I I I I

. i

-·-·-f- --=OCX)-+------~· ----·-1 . I 1 ; . I I . : . , - I

I I

3Z.

I I ' j I I - . I

~

I

sll 3~ sa

SAL /A// ~y V/t) I . ·-·· -·: . i !

. . .. ..... 1.. .. -

-· ;

I I . ___ j

\_ __

·- --i

3 I

1 .. i

I r . __ j

-~- -·- --. I

i

i ' I ___ _l ___ i__ ___ ____ ! _________ - - -

I

-l .......... --

i · I ... . ______ L_ -- ----· ..... __ j

FIGURE 6 ') 1

I

I I

--- -··· --··· -- j __ j ___ - !

I . I

·.. -· . - I

FIGURE 7

I -·---1-

l

I

I I

I I

E1 vr I I

-I I I

d ~

1

-i I

I I

1'0 I

.I FI GIIRE 8

~ . ~

!

i I

i 3'1 oM ~

SAt/NirYw1 1

-'·- - -·- . ____ l .. --·--_j

I

I . Gt1T£ .S.Jt.ltVITY .Zfr.:RcoNPAR!SOAI I : I I ____ _._- {INDI/CT/VE . SA/..//t/0-f'ETE/?-.BECI(HAN} ----~ ··--i ___ _

i i • n1m.d ~ SA-'rCS -: I __I

: --~3- -65f.-! I

: I I I

. I ..

' ; ! I

"' to

lb

·0

I

I

!• . --- - ,-1

I I

I I

-··-·-L ··--··· '· ------· -- .... : . . . . ! .. ! · -· t .- . _j

FlC~IJRE 9 )/1

l ·-

l·----• j ..

I -1 ..

I

I

I

~ I I

I I

-- _j

I

I

. .. . ... ' . -.. ... . -l. .. ---- -

I

. I ! J . .

..... · ---~~ - · - ·- · --··- ----·-·

;D ~:< J'f ~ .;'8 1/0

' SA~IIY/TY~~t) i I!

L I . I I . - -----·- -· _ ____ .. .. .. !. _______ ____ , ___ .!... -· _\ ___ .. J

FIGURE 10

-- - · • I---. t./0

: qAT£ $AuNlTY llNr..eRcoMP.aR;so..v I · · 1

-

I I ( I I l ' I ) I ~--· -·· · j · · .. : SAL.INOI1El7,E'R -GJo~ 6..f ---· --~ -1- _. -·i

I : I· . I ' I I II !

I ' I I ·-·--·-, ~

I

I

-· · ~ I

' • • I

'

! ... ~SAMPLES I -o--y SAHPLES I I

1 l l --62.8-- L. ! · -63/-:

I ' I I I i i I ' .lD I

I __ j __

I I

! ~ ~~~~--~--~~--~--~

i !· ... · 0 ~--__::,....,......___;~___;~1'-~<*--_;:,.,~

I I I ~I( ,. . .~IJ . <lfJ l

SALINITY~.I'J:) ' : i Zo

I i i ; -·· ·-.. ··i . .. . .. ····-,-· -·- - -- i -. 1/C I .....

I l I ;

-&9-

I I

; I I

• i

!

' 3~ 3~ 3tf w

SAL IN/TYV:>~)

. I FIGUREll

')(

I

</0

i I

3:1.

I

'

. I

"i I

!

. I l

• -- _ J

- ·-- ,--·-

.... .. -·--·-;

I

~ I

~ I I

• .,.J

:sJ I

'"'i I

I

1

I •• - _ J L.

I I GATE .5ALIN~TY ZNTERCONPARISON ! ! I

: ----~--- -I· j I • I I

__ /SA~INOI'IcrEA - GN 6S) I; __ --------l __ , ___ _ _ ( • I ; • I I

! I . I ' . I I . I i ·I I I . ! . I

.. X S4HPL£S I 0 YSAHPLES i

I I I :

lid . . . . 3.t ·-

I I

1/10'

i I I

I

• I ' . .

·-- l ______ _i __ _

I I l

3V __ -~ .... .68.

S~L/N 17Y(//i)-l' I I

'15-· I

-- -- I . ----- .

I I I I

__ j

1

: I 1

... 110 ... . ...•. .

FI GURE 12 27

; I_·-·--6¥9- : . I

10

11 I

I <t)

I .to

I I ' .

I ~

I I .. - .. - r ..

I j I

i I

IOI----'""--~---'!:>o-~~.._-+

;%. _ _ 3¥ _ _ J'L 3~ I SAL/N/'TY~/1) I I

·-• .. - - --- -- --· · --- . I

I ' I I

20

' . j ~ j~ . : ~ ... J

~

I

I

- . ' ' , I

' ' I I I

~ -0 · 1-------~ I

0-- .I ... -·--1

·- oi -l

I

I .

. I : ! I --- - -o - ·---- -~ ·- - o - 0 ~- o· - · --1 I. I

0 I

0

• I I I . . l 0 ; • ; I . : --:. I

- . ! ·--- ~ 0 o• 0 •• 11: ; - ·- I ·-· - --i--·- . . :- . -~~ - .. -.-----1- . __ 1.. -- ~ j ' : i I I

I . ! -----; I

l : i I : I ; 'I . I ! I : 0 0 I i I . ~ ! i : I

--~-: -~-r- -·-1 · -- - : --- __ _j__ -

1--+- --

1

----i·--1

1-0

-

0

--T 0

-

0

I l I 0 -~ ·- l . 0 ... -

I ' i I ! o I ' i i I : I .

1° I 1 • o

' _j I __ ;----·-r __ ,_ . 0 __

1 ______

0• _ _ • io --·-· __ , --·-_____ . ___ 0 : ___ -;----- .-0-0 _ ••

I I ; i l

L_ _ l_

I I I I

I.---- . ----· I I

_ .. l - .

FIGIIP.E 13 .... 0

I

! ..

I I

_____ !_ __

.

_j

NOMit~AI

SAMPLE1

NUMBE OF SAt

SET

24

4

4

lL

1•

,. (.

I

- ·'

- ·-· --1

I - - _j

Table 10 indicates the average and standard deviations for the titration method and each instrument. The data source i s from measurements obtai ned from several sets of water samples that include X and Y sets. For example, the deviations for titration were derived from 144 measurements (24 sets x 6 samples per set = 144 measurements). Guildline electrode salinometer, 24 measurements, etc.

TABLE 10. Da ta deviations of NOIC's PCC and other systems to the PCC.

NOMINAL SAL. {ppt) 32.0 33.5 35.0 36.5 38 .0

SAMPLE GROUP NO . X05 Yll X04 YlO Y12 X03 Y09 X02 X06 Y08 X01 Y07

NUMBER OF SAt.WLE EQUIPMENT

SET

24 Titration (ppm) (ppm) (ppm) (ppm) (opm} Avg. Dev. -19 -9 -2 -24 -5 -24 5 -8 -7 -10 -6 2 Std. Oev. 24 26 13 23 20 14 17 26 25 27 26 38

4 Guildl ine Avg. Oev. 1 2 0 2 1 1 2 2 0 2 1 0 Std. Oev. 2 2 2 1 0 1 1 4 1 1 2 1

4 WHOI Avg. Dev. -3 -1 -1 -1 0 -4 0 -4 -1 0 -4 - 1 Std. Dev. 3 3 2 3 0 5 3 5 2 3 5 4

18 PCC Avg. Oev. 0 0 0 0 0 0 0 0 0 0 0 0 Std. Dev. 1 1 1 1 1 1 2 1 1 1 1 1

14 P1essey Avg . Dev . 7 6 6 4 4 4 0 1 -4 1 -1 -5 Std. Oev . 6 6 4 6 5 5 4 9 4 5 5 4

14 Beckman Avg. Dev. 5 7 3 -3 -1 -2 -6 -7 -9 -12 -13 -15 Std. Dev . 11 11 11 11 11 17 8 16 15 8 17 11

8 Auto Lab Avg . Dev. 16 10 4 17 14 1 4 1 12 0 -1 -3 Std. Oev. 21 14 4 25 24 5 6 4 23 8 5 8

26 GM65 Avg. Dev . 13 15 13 11 7 -3 4 5 1 6 -2 1 Std. Dev. 17 14 17 16 16 15 10 15 12 13 16 17

29

The 108 measurements taken by NOIC on its PCC before , during and after the GATE project had a standard deviat ion of 1.1 ppm as shown in Table ll . The randomly sel ected samples, tested by other laboratories with simil iar thermostated equipment, had a s tandard deviation of 1.1 ppm . The collaborative measurements by other laboratories with titration and with the use of non-thermostated equipment served the purpose of demonstrating a typical performance for those measurement types.

TABLE 11 . Analys is of t he quality control data .

QUANTITY STANDARD OF DATA OBSERVED DEVIATION

MEASUREMENT GROUP MEASUREMENTS WITHIN + 1 0 ppm OF GROUP (ppm)

NOIC Inhouse Control 108 100 1.1

Collaborative Titration 12 50 18.6

Collaborative Non-Thermostated 24 46 13.0

Collaborative Thermostated 26 100 1.1

V. CONCLUSIONS

The GATE Salinity Intercomparison results from the field data indi cates that:

{l) The titration method for measuring salinities within +10 ppm is unrel iable and probably beyond its capabilities. Many of the data poi nts were very erratic and trends in the overall performances could not be calculated.

(2) The electrode-thermostated type salinometers performed superior to other types of methods and equipment in terms of data accuracy. Its maintainability is unknown at this time due to limited utilization among scientists.

(3) The conventional inductive-non-thermostated type salinometer exhibited performances that have become typical with this instrument, such as: generally satisfactory when used with salinities close to 35 ppt; a negative error slope through 35 ppt where positive differences occur at lower salinities than 35 ppt and negative di fferences at higher than 35 ppt salinities; and frequent standardi­zation with standard seawater and adjustments can produce improved performances.

30

aboal X se· at t ment equi data ment timE

the rec cal mea set mec sh1

th co an on ir GJ Sl s v d

· a t

0 ON (ppm)

In revie\·1, the opr~ r'ational schE.:llle for mcasut"ing the samples aboard the GATE vessel:, v.ras to detennine the salinity of the X set of samp 1 es befc.re the star t of a GATE phase and of the Y set at the conclusion of a phase . Hopefully , the resultant measur e­ments would produce a stability and/or drift tr·end \'dth the equipment . Discounting the titration method because of errati c data , 55 percent of the X and Y set measurements indicated equip­ment dri ft equal t o or greater than 10 ppm for an approixmate time di fferential of 30 days.

I t is not possible to extract meaningful conclusions about the equipment performance abroad the vessels . Information was not recei ved or instructions given to the vessels pertaining to cal ibration procedures, operational instructions on the salinity measuring equipment, traini ng of the operators , etc. The sali nity sets v1ere delivered to the vessels with minimal instructions: measure the samples, record the appropri ate data on the da ta sheets and on completion, forward them to NOIC.

There is no reason to doubt the stability of the samples and the va 1 i dity of their va 1 ues obtai ned vd th the PCC. A 11 qua 1 i ty control checks conducted on the PCC and the samples before, during and after the GATE exerci se, including intercomparison meas urements on similar type of equipment \'lith reputable laboratories, resulted in approximately 1 ppm standard deviation. A report from one of the GATE vessels t hat a white, crystalline material appeared in the sarol e . could not be dupl i cated in NOIC's t esti ng laboratory. Samp les of the same solutions , that were forwarded to the GATE vessels, were exposed to a broad range of harsh environmental con­ditions with negative results. No white crystalline material or

· any other apparent deterioration of the water was observed during the tests .

ACKNOWLE DGEMENT

The author wishes to express his gratitude and special thanks to Ed\'lard Snead, James Sprenke and all those members of NOI C's Metrology Divisi on for their generous support given in the car ryi ng out of thi s GATE Salinity Intercomparison Project . The col laborative meas urements performed by the U. S. Naval Oceanographic Office (J. Sul livan), Woods Hole Oceanographic Institution (K. Schleicker), Canada's Bedford Institut e of Oceano~raphy ( J. Bed l em) and Inst itute of Oceanographic Sciences {F. Culkin) in support of this project are also greatly apprec iated.

31

DATA I

CODE

601 602 603 604 605 606

I! 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 ,___

DATA DATES OF INSTRUMENT/METHOD J CODE f~EASURE COUNTRY VESSEL

601 2 APR-5 APR USA NOIC (Hash} PCC 602 2 APR-5 APR USA NOIC (Was h) PCC 603 7 JUN-7 JUN USA NOIC (\~ash) PCC 604 15 JUL-16 JUL USA NOIC (\~ash} PCC 605 7 AUG-7 AUG USA NOIC (Wash) PCC 606 20 AUG-20 AUG USA NOIC (Hash) PCC (250) 607 28 AUG-28 AUG USA NOIC (Wash) PCC (150) 608 17 SEP-17 SEP USA NOIC (Hash) PCC (250) 609 1 OCT-4 OCT USA NOIC (Hash) PCC 610 27 NOV-27 NOV USA NOIC (Wash) PCC (ENV TEST) 511 11 JUN-11 JUN CANADA AOL (BlO) Guildline 8400 612 4 SEP-4 SEP CANADA AOL (61 0) Gui1d1ine 8400 613 26 JUN-27 JUN USA NAVOCEANO Beckman RS-78 614 27 JUN-27 JUN USA NAVOCEANO Titration (Knudsen) 615 22 AUG-22 AUG USA WHOI (Schleicher} WHOI #3 516 25 FEB-25 FEB UK I OS ( F . Cu H i n ) Auto lab 601MKIII 617 17 JUL-17 JUL USA DALLAS (716) Beckman RS-7B 618 16 JUN-1 JUL USA ATLANTIS II WHO! #2 ' 619 12 JUL-29 JUL USA ATLANTIS II WHOI #2 620 18 JUN-19 JUL USA GILLIS$ Beckman RS-7C 621 8 AUG-18 AUG USA GILLIS$ Beckman RS-7C 622 30 JUN-20 AUG USA TRIDENT P1essey 6230'~ 623 0 JUN-27 JUL USA OCEANOGRAPHER Beckman ns-7B 624 17 JUL-20 SEP USA OCEANOGRAPHER Bec!Qnan RS -78 625 1 JUL-23 JUL USA RESEARCHER Plessey 6230 626 28 AUG-10 OCT USA RESEARCHER P1essey 6230 627 18 AUG-22 AUG USA H.J.H . FAY Auto Lab 601MKIII 628 26 AUG-22 SE"P USSR PASSAT GH-65 629 3 SEP-20 SEP USSR M. LOMOROSOV GM-65 630 25 JUL-24 AUG USSR OCEAN GM-65 631 29 AUG-21 SEP USSR OCEAN GM-65 632 30 AUG-22 SEP USSR ADADEMIK KURCHATOV Gt1-65 633 25 JUN-16 JUL USSR MUSSON Gr4-65 o34 17 AUG-19 SEP USSR ~1USSON Gl~-65 635 25 JUN-16 JUL USSR MUSSON Titration 636 17 AUG-19 SEP USSR MUSSON Titration 6J7 27 JUL-15 AUG USSR PRIBOY Ti tration 638 28 AUG-19 SEP USSR PRIBOY Titration 639 28 JUN-16 JUL USSR AKADEMIK KOROLOV Titration 640 28 JUN-16 JUL USSR AKADEMIK KOROLOV Auto-T1trat1on

lll ll~lllirnlli~l ~~I M~~lll ~~~~11111~~~~11 1 3 1222 00516 198 2

ATTACHt~ENT 1 (Continued)

DATA CODE KEY

DATA DATES OF CODE MEASURE COUNTRY VESSEL INSTRUMENT /METHOD

641 28 JUL-14 AUG USSR AKADEMIK KOROLOV Titration 642 28 JUL-1 4 AUG USSR AKADENIK KOROLOV Auto-Titrati on 643 30 AUG-17 SEP USSR AKADEt4I K KOROLOV Titration 644 30 AUG-1 7 SEP USSR AKADEf~I K KOROLOV Auto-Titration 645 24 J UN-29 JUL USSR ERNST KRENKEL GM-65 646 20 AUG-19 SEP USSR ERNST KRENKEL Gr4-65 647 24 J UN-29 JUL USSR ERNST KRENKEL Titration 648 20 AUG-19 SEP USSR ERNST KRENKEL Titration 649 24 JUN-16 JUL USSR PORYV GM-65 650 28 JUL-16 AUG USSR PORYV GM-65 651 29 AUG-19 SEP USSR PORYV GM-65 652 0 SEP-15 SEP- USSR DEZHNEV GM-65 653 29 JUN-15 JUL FRANCE JEAN CHARCOT Beckman RS-78 654 1 SEP-23 SEP FRANCE CAPRI CORNE Plessey 6230 655 26 JUN-19 SEP BRAZIL ALMIRANTE SALDANHA Plessey 6230N 656 27 JUL-17 AUG BRAZIL AUH RANTE SALDANHA Plessey 6230N 657 31 JAN-31 JAN FRG METEOR PLANET DOHRN Beckman RS-7B 658 28 JUN-17 SEP NETHERLANDS ONVERSAAGD Auto Lab 601MKI II 659 2 SEP-17 SEP UK DISCOVERY Plessey 6230 660 27 AUG-28 AUG UK DISCOVERY Auto Lab 601MKII I 661 1 JUL-30 JUL CANADA QUA ORA Guildline 8400 #2 662 30 JUL-22 SEP CANADA QUADRA Guild1ine 8400 #2 663 28 AUG-28 1\UG GDR A. V. HUt1BOLDT Auto Lab 601MKIII